Epithelial to mesenchymal transition (EMT) has been enthusiastically proposed as an essential mechanism for tumor metastasis, since the EMT-associated features such as migration, invasion, resistance to apoptosis and stemness properties, adequately meet the requirements for metastasis. Taken the challenges of tracing the EMT process in vivo, we developed a strategy of using a mesenchymal-specific Cre-mediated switch of fluorescent markers in a multiple-transgenic mouse (MMTV-PyMT:Fsp1-Cre:Rosa26mT/mG, Tri-PyMT). Surprisingly, we found that lung metastases were predominantly composed of pre-EMT RFP+ tumor cells exhibiting epithelial phenotypes under normal conditions. Importantly, the post-EMT tumor cells did exhibit resistance to chemotherapy, significantly contributed to recurrent lung metastases after chemotherapy. These findings pointed to the complexity of EMT contributions in tumor progression and revived vigorous discussions in the community. Given the transient, reversible and dynamic natures of the EMT process, and concerns about the Tri-PyMT model, we proposed new lineage tracing models to study the roles of EMT in metastasis and chemoresistance. Aim 1. To explore the contributions of EMT mechanism in tumor metastasis and chemoresistance by using Snail-CreERT2 mediated EMT lineage tracing model. The Fsp1-Cre mediated Tri-PyMT model may not be sufficient to report all EMT events. Metastatic cells could undergo EMT by activating distinct EMT transcription factors (TFs) such as Snail. Therefore, we have established a Snail-CreERT2?mediated EMT lineage tracing model. In-depth analyses will be performed to clarify the roles of EMT in metastasis and chemoresistance with this model. Aim 2. To explore the dynamic changes of EMT statuses in human breast cancer metastasis and chemoresistance. EMT reporter cell lines (MDA-MB-231:Vim/RFP and BT-474:ECAD/GFP cells) carry knockin fluorescent tags within EMT marker genes (Vimentin and E-cadherin, respectively). Unlike the Cre-mediated models, the fluorescence expression in these cells is quantitative and reversible, which allows us to analyze the dynamic changes of EMT status with and without chemotherapy in human breast cancer models. Aim 3. To assess the evolutionary lineages of metastasis-initiating cells and the involvement of EMT mechanism via genetic barcoding models. In addition to using EMT markers, we will genetically barcode the Tri-PyMT cells using the homing-CRISPR technique. This model will allow us to depict the evolutionary trajectories from primary tumor cells to individual metastases, and determine the origins of the metastasis (pre- EMT vs. post-EMT cells). Further, we will develop genetic barcoding mice (MARC:CRISPR:PyMT) for lineage tracing of spontaneous metastatic cells and assessing the contributions of the EMT program to metastasis. Impact: Resolving the controversies in the field will not only improve our mechanistic understanding of tumor metastasis but also provide novel targets/opportunities in combatting the deadly disease.